计算机科学 ›› 2014, Vol. 41 ›› Issue (5): 235-238.doi: 10.11896/j.issn.1002-137X.2014.05.049

• 人工智能 • 上一篇    下一篇

基于忠诚度的多智能体联盟效用分配策略

曹义亲,张贞,黄晓生   

  1. 华东交通大学软件学院 南昌330013;华东交通大学软件学院 南昌330013;华东交通大学信息工程学院 南昌330013
  • 出版日期:2018-11-14 发布日期:2018-11-14
  • 基金资助:
    本文受江西省教育厅科技项目(GJJ12305,GJJ12306),江西省科技支撑计划项目(20123BBE50093),江西省研究生创新专项资金项目(YC2012-X018),国家自然科学基金项目(61365008,61163056)资助

Multi-agent System Coalition Utility Allocation Strategy Based on Loyalty

CAO Yi-qin,ZHANG Zhen and HUANG Xiao-sheng   

  • Online:2018-11-14 Published:2018-11-14

摘要: 为增强多智能体系统中联盟成员效用划分的合理性,促使agent形成稳定的全局最优联盟,提出了一种基于忠诚度的多智能体联盟效用分配策略。新策略引入agent忠诚度的概念,根据agent每次参与联盟后是否服务到任务结束来对agent忠诚度进行评价,从而表示各agent的忠诚度大小,将忠诚度与各agent完成任务的能力相结合,共同决定联盟内各agent效用如何划分。理论分析和实验表明,新策略提高了对联盟效用分配的合理性,较好地满足了联盟形成机制的全局最优性、强稳定性、时效性、简单性等要求。

关键词: 多智能体系统,联盟,效用分配,忠诚度

Abstract: In order to enhance the rationality about the coalition utility allocation among agents and the stability about forming the global optimal coalition in multi-agent system,this paper proposed a coalition formation strategy based on loyalty.The new strategy introduces the conception of agent’s loyalty,determines if one agent leaves the coalition before finishing the task or not during agent taking part in a coalition every time,and consequently evaluates the agent’s loyalty.At the same time,the new strategy decides the coalition utility allocation by means of combining each agent’s loyalty and their ability to finish the task.Theoretical analysis and experiment results show that the novel strategy can improve the justice of allocation for utility and achieve a global optimal solution,which is stable,speedy and simple.

Key words: Multi-agent system (MAS),Coalition,Utility allocation,Loyalty

[1] 胡立山,石纯一,李少芳.给定限界的势结构分组与联盟结构生成[J].计算机学报,2012,5(12):2618-2624
[2] 刘惊雷,张伟,童向荣,等.一种O(2.983n)时间复杂度的最优联盟结构生成算法[J].软件学报,2011,2(5):938-950
[3] 裘杭萍,覃垚,胡汭,等.多Agent系统中基于改进合同网模型的任务分配研究[J].计算机科学,2012,9(6A):279-281
[4] Shapley L S.A value for n-person games[M]∥Roth A E.The Shapley Value Cambridge:Cambridge University Press,1995:31-40
[5] 罗翊,石纯一.Agent协作求解中形成联盟的行为策略[J].计算机学报,1997,0(11):961-965
[6] 蒋建国,夏娜,于春华,等.基于能力向量发挥率和拍卖的联盟形成策略[J].电子学报,2004,2(12A):215-217
[7] 夏娜,蒋建国,于春华,等.一种基于利益均衡的联盟形成策略[J].控制与决策,2005,0(12):71-75
[8] 张国富.基于群体智能的复杂联盟机制研究[D].合肥:合肥工业大学,2008:51-66
[9] 李剑,景博,杨义先.一种基于奖励机制的agent联盟形成策略[J].电子学报,2008,7(12A):71-75
[10] 问翠梅.多Agent系统中联盟形成问题的研究[D].兰州:兰州大学,2009:34-38
[11] Huynh T D,Jennings N R,Shadbolt N R.FIRE:An integrated trust and reputation model for open multi-agent systems[C]∥Proc.16th European Conerence on Aritificial Intelligence.Valencia,Spain,2004:18-22
[12] Zoltkin G,Rosenschein J S.Coalition,Cryptography and Stability:Mechanisms for Coalition Formation in Task Oriented Domains[C]∥Proc of the National Conf on Artificial Intelligence.Seattle:AAAI Press,1994:432-437

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!